more fixes

Signed-off-by: Pratik Mankawde <3397372+pratikmankawde@users.noreply.github.com>

.github/workflows/reusable-build-test-config.yml

@@ -206,7 +206,7 @@ jobs:
         if: ${{ !inputs.build_only && env.SANITIZERS_ENABLED == 'true' }}
         run: |
           echo "ASAN_OPTIONS=print_stacktrace=1:detect_container_overflow=0:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/asan.supp" >> ${GITHUB_ENV}
-          echo "TSAN_OPTIONS=second_deadlock_stack=1:halt_on_error=0" >> ${GITHUB_ENV}
+          echo "TSAN_OPTIONS=second_deadlock_stack=1:halt_on_error=0:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/tsan.supp" >> ${GITHUB_ENV}
           echo "UBSAN_OPTIONS=suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/ubsan.supp" >> ${GITHUB_ENV}
           echo "LSAN_OPTIONS=suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/lsan.supp" >> ${GITHUB_ENV}
 

cspell.config.yaml

@@ -96,6 +96,7 @@ words:
   - doxyfile
   - dxrpl
   - endmacro
+  - eventfd
   - exceptioned
   - Falco
   - finalizers

sanitizers/suppressions/tsan.supp

@@ -1,101 +1,45 @@
-# The idea is to empty this file gradually by fixing the underlying issues and removing suppresions.
+# TSAN suppression file for rippled.
+# Only suppress issues in third-party libraries and TSAN's own instrumentation.
+# Races in rippled's own code should be fixed, not suppressed.
 
-# Suppress race in Boost ASIO scheduler detected by GCC-15
-# This is a false positive in Boost's internal pipe() synchronization
+# Boost ASIO / Boost Context false positives
+# These are internal to Boost's reactor, fiber context switching, and memory management.
 race:boost/asio/
 race:boost/context/
 race:boost/asio/executor.hpp
 race:boost::asio
-
-# Suppress tsan related issues in rippled code.
-race:src/libxrpl/basics/make_SSLContext.cpp
-race:src/libxrpl/basics/Number.cpp
-race:src/libxrpl/json/json_value.cpp
-race:src/libxrpl/json/to_string.cpp
-race:src/libxrpl/ledger/OpenView.cpp
-race:src/libxrpl/net/HTTPClient.cpp
-race:src/libxrpl/nodestore/backend/NuDBFactory.cpp
-race:src/libxrpl/protocol/InnerObjectFormats.cpp
-race:src/libxrpl/protocol/STParsedJSON.cpp
-race:src/libxrpl/resource/ResourceManager.cpp
-race:src/test/app/Flow_test.cpp
-race:src/test/app/LedgerReplay_test.cpp
-race:src/test/app/NFToken_test.cpp
-race:src/test/app/Offer_test.cpp
-race:src/test/app/ValidatorSite_test.cpp
-race:src/test/consensus/NegativeUNL_test.cpp
-race:src/test/jtx/impl/Env.cpp
-race:src/test/jtx/impl/JSONRPCClient.cpp
-race:src/test/jtx/impl/pay.cpp
-race:src/test/jtx/impl/token.cpp
-race:src/test/rpc/Book_test.cpp
-race:src/xrpld/app/ledger/detail/InboundTransactions.cpp
-race:src/xrpld/app/main/Application.cpp
-race:src/xrpld/app/main/GRPCServer.cpp
-race:src/xrpld/app/misc/detail/AmendmentTable.cpp
-race:src/xrpld/app/misc/FeeVoteImpl.cpp
-race:src/xrpld/app/rdb/detail/Wallet.cpp
-race:src/xrpld/overlay/detail/OverlayImpl.cpp
-race:src/xrpld/peerfinder/detail/PeerfinderManager.cpp
-race:src/xrpld/peerfinder/detail/SourceStrings.cpp
-race:src/xrpld/rpc/detail/ServerHandler.cpp
-race:xrpl/server/detail/Door.h
-race:xrpl/server/detail/Spawn.h
-race:xrpl/server/detail/ServerImpl.h
-race:xrpl/nodestore/detail/DatabaseNodeImp.h
-race:src/libxrpl/beast/utility/beast_Journal.cpp
-race:src/test/beast/LexicalCast_test.cpp
-race:ripple::ServerHandler
-
-# More suppressions in external library code.
-race:crtstuff.c
-race:pipe
-
-# Deadlock / lock-order-inversion suppressions
-# Note: GCC's TSAN may not fully support all deadlock suppression patterns
-deadlock:src/libxrpl/beast/utility/beast_Journal.cpp
-deadlock:src/libxrpl/beast/utility/beast_PropertyStream.cpp
-deadlock:src/test/beast/beast_PropertyStream_test.cpp
-deadlock:src/xrpld/core/detail/Workers.cpp
-deadlock:src/xrpld/app/misc/detail/Manifest.cpp
-deadlock:src/xrpld/app/misc/detail/ValidatorList.cpp
-deadlock:src/xrpld/app/misc/detail/ValidatorSite.cpp
-
-signal:src/libxrpl/beast/utility/beast_Journal.cpp
-signal:src/xrpld/core/detail/Workers.cpp
-signal:src/xrpld/core/JobQueue.cpp
-signal:ripple::Workers::Worker
-
-# Aggressive suppressing of deadlock tsan errors
-deadlock:pthread_create
-deadlock:pthread_rwlock_rdlock
-deadlock:boost::asio
-
-# Suppress SEGV crashes in TSAN itself during stringbuf operations
-# This appears to be a GCC-15 TSAN instrumentation issue with basic_stringbuf::str()
-# Commonly triggered in beast::Journal::ScopedStream destructor
-signal:std::__cxx11::basic_stringbuf
-signal:basic_stringbuf
-signal:basic_ostringstream
-
-called_from_lib:libclang_rt
-race:ostreambuf_iterator
-race:basic_ostream
-
-# Suppress SEGV in Boost ASIO memory allocation with GCC-15 TSAN
-signal:boost::asio::aligned_new
-signal:boost::asio::detail::memory
-
-# Suppress SEGV in execute_native_thread_routine
-signal:execute_native_thread_routine
-
-# Suppress data race in Boost Context fiber management
-# This is a false positive in Boost's exception state management during fiber context switching
-race:__cxxabiv1::manage_exception_state
 race:boost::context::fiber::resume
 race:boost::asio::detail::spawned_fiber_thread
 race:boost::asio::detail::spawned_fiber_thread::suspend_with
 race:boost::asio::detail::spawned_fiber_thread::destroy
-
-# Suppress data race in __tsan_memcpy called from Boost fiber operations
+race:__cxxabiv1::manage_exception_state
 race:__tsan_memcpy
+
+# TSAN's own syscall interceptors (false positives from fd tracking)
+race:crtstuff.c
+race:pipe
+race:epoll_ctl
+race:epoll_create
+race:closedir
+race:close
+race:socket
+race:accept
+race:eventfd
+
+# C++ standard library internals
+race:ostreambuf_iterator
+race:basic_ostream
+called_from_lib:libclang_rt
+
+# Deadlock false positives in Boost and threading primitives
+deadlock:pthread_create
+deadlock:pthread_rwlock_rdlock
+deadlock:boost::asio
+
+# Signal/crash suppressions for GCC TSAN instrumentation issues
+signal:std::__cxx11::basic_stringbuf
+signal:basic_stringbuf
+signal:basic_ostringstream
+signal:boost::asio::aligned_new
+signal:boost::asio::detail::memory
+signal:execute_native_thread_routine

src/xrpld/app/main/Application.cpp

@@ -1035,7 +1035,13 @@ private:
 bool
 ApplicationImp::setup(boost::program_options::variables_map const& cmdline)
 {
-    startIOThreads();
+    // NOTE: IO threads are intentionally NOT started here.
+    // All setup work is non-blocking (it enqueues async work to the
+    // io_context but doesn't require it to be processed yet).
+    // IO threads are started at the beginning of start() to avoid
+    // data races between the main thread doing setup and IO threads
+    // processing enqueued work concurrently.
+
     m_resourceManager->start();
     m_nodeStore->startReadThreads();
 
@@ -1370,6 +1376,11 @@ ApplicationImp::start(bool withTimers)
 {
     JLOG(m_journal.info()) << "Application starting. Version is " << BuildInfo::getVersionString();
 
+    // Start IO threads now that all setup is complete.
+    // This must happen before starting subsystems that post work
+    // to the io_context (resolver, overlay, etc.).
+    startIOThreads();
+
     if (withTimers)
     {
         setSweepTimer();